Machine Learning Regression Algorithm For Sales Prediction

Team: Eduardo Villalpando Mello, Juan Carlos Garfias Tovar & Luis Alberto Fernández

Graph

Data

ID Location Product Date Temp Mean Temp Max Temp Min Sunshine Quantity Event Price ($) Predicted Sales Quantity

How it works

The machine learning model utilizes the libraries pandas, numpy and sklearn to find a correlation between the given data.


Steps

1) The algorithm checks for the elements that lack data then it proccedes with imputation using mean values.

2) The algorithm changes the event values to binary in order to change the column to categorical data.

3) The algorithm changes the location to a weight by replacing it to the categorical mean.

4) The date is replaced to the day of the week due to the correlation with beer consumption.

5) The dataset is converted into a matrix.

6) A linear multivariable regression is applied on the matrix with location, product, date, temp_mean, sunshine, event and price.

7) The prediction values are stored and append to the dataset.

8) The dataset is converted into a csv.